Shoe and clothing product defect positioning method and system based on multi-step reasoning and mapping knowledge domain

Abstract

The invention provides a shoe and clothing product defect positioning method and system based on multi-step reasoning and a knowledge graph, and the method comprises the steps: constructing a three-dimensional knowledge graph comprising three node dimensions, and the node dimensions comprise a material dimension, a process dimension and a quality defect dimension; acquiring data of each node dimension; and the inference engine performs reverse inference along the quality defect dimension, the process dimension, the material dimension and the environment dimension according to the quality defect real-time data acquired by the sensor, generates a multi-step inference path, and determines the cause of the defect. According to the method, the influence of the temperature and the humidity on the ductility of the leather is corrected through the rule compensation library, so that the phenomenon of stitch length offset caused by the influence of the temperature and the humidity when a craftsman sews the leather is avoided, and the phenomena of leather breakage or loose connection and leather quality reduction are finally caused.

Description

Technical Field

[0001] The present invention relates to the field of knowledge graph technology, and in particular to a method and system for locating defects in footwear and clothing products based on multi-step reasoning and knowledge graphs. Background Art

[0002] In the footwear and apparel manufacturing industry, especially in processes involving leather processing, operations are not completely immune to environmental interference. Numerous practical observations have revealed that the same craftsman's stitch length control can exhibit systematic deviations when sewing in different weather conditions. For example, in humid weather, the ductility of vegetable-tanned leather increases, and the customary stitch length becomes too dense on the stretched leather, resulting in tight seams that are prone to breakage during subsequent wear due to further deformation of the leather. In dry weather, the ductility of chrome-tanned leather decreases, and maintaining the same stitch length can result in sparse stitches, making it difficult to securely connect leather components and reducing the overall quality of the footwear and apparel. Summary of the Invention

[0003] Based on the above technical problems, the present invention proposes a method and system for locating defects in footwear and clothing products based on multi-step reasoning and knowledge graphs, and the technical solutions adopted are as follows: A method for locating defects in footwear and clothing products based on multi-step reasoning and knowledge graphs, the method comprising: S1: Construct a three-dimensional knowledge graph including three node dimensions, wherein the node dimensions include material dimension, process dimension, and quality defect dimension; S2: Get the data of each node dimension; S3: The reasoning engine uses the real-time quality defect data obtained by sensors to reversely reason along the quality defect dimension, process dimension, material dimension, and environmental dimension to generate a multi-step reasoning path and determine the cause of the defect.

[0004] Preferably, the node dimensions of S1 include environment dimension, material dimension, process dimension and quality defect dimension, specifically including: Said material dimensions include, leather type, leather thickness and leather surface characteristics; The process dimensions include the theoretical value of needle length and presser foot force; The quality defect dimensions include seam cracking, skewed stitches and uneven stitch length.

[0005] Preferably, the leather types include vegetable tanned leather and chrome tanned leather.

[0006] Preferably, the three-dimensional knowledge graph of S1 also includes relationship edge attributes, and the relationship variable attributes include the influence of temperature and humidity on materials, the needle distance compensation of the process on materials, and the pressure foot force.

[0007] Preferably, in addition to acquiring data of each node dimension, S2 also acquires the following data, specifically including: The ambient temperature and humidity are obtained through the temperature and humidity sensor, and the real-time pressure foot force is obtained through the electric pressure sensor.

[0008] Preferably, the reasoning engine of S3 further includes a compensation rule base, and compensates the needle distance according to the rules in the compensation rule base, specifically including: Determine the compensation method for stitch length according to leather type and humidity; When the leather type is vegetable tanned leather and the humidity is greater than the first compensation threshold, the first compensation method is used for compensation; When the leather type is chrome tanned leather and the humidity is less than the second compensation threshold, the second compensation method is used for compensation.

[0009] Preferably, the compensation method specifically includes: The first compensation method includes: when the leather type is vegetable tanned leather and the humidity is greater than the first compensation threshold, the stitch length compensation amount is adjusted in a decreasing direction for every 1% increase in humidity; at the same time, when the temperature exceeds 25°C, the stitch length compensation amount is adjusted in a decreasing direction for every 1°C increase in humidity. The final stitch length compensation amount is the sum of the two parts; The second compensation method includes adjusting the stitch length compensation amount in an increasing direction for every 1% decrease in humidity when the leather type is chrome-tanned leather and the humidity is less than a second compensation threshold.

[0010] Preferably, after the compensation rule library compensates for the stitch length, the adjusted presser foot force is obtained according to the stitch length compensation amount, and the adjusted presser foot force is displayed on the visual interface. The system obtains and displays the real-time presser foot force on the visual interface based on the electric pressure sensor, and adjusts the real-time presser foot force to be consistent with the adjusted presser foot force.

[0011] Preferably, the S3 specifically includes: Acquire real-time data on product quality defects, match the real-time data with the quality defect dimensions in the three-dimensional knowledge graph, confirm the corresponding quality defect nodes, and use the corresponding quality defect nodes as the starting point for reasoning. Based on the preset relationship edges, reverse reasoning is performed along the quality defect dimension, process dimension, material dimension, and environmental dimension to generate a multi-step reasoning path and ultimately determine the cause of the defect.

[0012] A shoe and clothing product defect location system based on multi-step reasoning and knowledge graph, characterized in that the system includes: Multi-dimensional knowledge graph construction system: constructs a three-dimensional knowledge graph with three node dimensions, including material dimension, process dimension and quality defect dimension; Full-dimensional data acquisition system: obtain data of each node dimension; Defect tracing reasoning system: The reasoning engine uses the real-time quality defect data obtained by sensors to reversely reason along the quality defect dimension, process dimension, material dimension, and environmental dimension to generate a multi-step reasoning path and determine the cause of the defect.

[0013] The present invention has the following beneficial effects: The present invention uses a rule-based compensation library to compensate for the effects of temperature and humidity on leather ductility, thereby preventing stitch deviation caused by temperature and humidity when sewing leather, which ultimately leads to leather breakage or loose connections, reducing leather quality. Furthermore, the present invention's full-dimensional data acquisition system acquires dimensional data from each node in real time. The inference engine uses this data, combined with a knowledge graph and compensation rule library, to perform reasoning and decision-making, achieving an automated process from quality defect data to cause location and process parameter adjustment, thereby improving production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 This is a method for locating defects in footwear and clothing products based on multi-step reasoning and knowledge graphs as described in the present invention. DETAILED DESCRIPTION

[0015] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0016] One embodiment of the present invention provides a method for locating defects in footwear and clothing products based on multi-step reasoning and knowledge graphs, the method comprising: S1: Construct a three-dimensional knowledge graph including three node dimensions, wherein the node dimensions include material dimension, process dimension, and quality defect dimension; S2: Get the data of each node dimension; S3: The reasoning engine uses the real-time quality defect data obtained by sensors to reversely reason along the quality defect dimension, process dimension, material dimension, and environmental dimension to generate a multi-step reasoning path and determine the cause of the defect.

[0017] The working principle and effect of the above technical solution are as follows: First, in stage S1, a three-dimensional knowledge graph is constructed, including material, process, and quality defect dimensions. The attributes of the relationship edges are also determined, thereby establishing a correlation framework between the dimensions. Next, in stage S2, data for each node dimension is acquired. Finally, in stage S3, the inference engine matches the real-time sensor-generated quality defect data of footwear and apparel products with the quality defect dimensions in the three-dimensional knowledge graph, identifying the corresponding quality defect node and using it as the starting point for inference. Subsequently, based on the pre-set relationship edges, reverse inference is performed along the quality defect, process, material, and environmental dimensions. During the inference process, the relationships between the dimensions established in the knowledge graph are leveraged to gradually trace the causes of the quality defects. This method considers and avoids the possibility that environmental factors may alter leather properties and thus affect product quality, thereby enhancing the environmental adaptability of the production process.

[0018] In one embodiment of the present invention, the node dimensions of S1 include an environment dimension, a material dimension, a process dimension, and a quality defect dimension, specifically including: Said material dimensions include, leather type, leather thickness and leather surface characteristics; The process dimensions include the theoretical value of needle length and presser foot force; The quality defect dimensions include seam cracking, skewed stitches and uneven stitch length.

[0019] The working principle and effect of the above technical solution are as follows: in the reasoning process of shoe and clothing defect positioning, the leather types in the material dimension have different sensitivities to temperature and humidity due to differences in their internal fiber structure and chemical properties. For example, vegetable tanned leather may be more likely to absorb water and soften when the humidity increases, causing the fiber spacing to change, affecting the stitch length; leather thickness is different, and the degree of deformation under temperature and humidity changes is different. Thicker leather expands more obviously when the humidity is high, requiring the stitch length to be adjusted accordingly; leather surface properties, such as smoothness and roughness, will change the friction between the suture and the leather under the influence of temperature and humidity, thereby affecting the stitch length setting. In the process dimension, although the theoretical value of the stitch length is a basic reference, after the temperature and humidity cause material changes, the actual stitch length needs to deviate from the theoretical value. For example, when the humidity is high, in order to prevent the suture from being too tight and breaking or the leather from being overstretched, the stitch length needs to be appropriately increased; the presser foot force is also adjusted due to the material properties changed by temperature and humidity. When the material becomes soft, the presser foot force needs to be reduced; The characteristics and quality of materials are potential factors leading to quality defects. The physical and chemical properties of materials, such as the flexibility, strength, and abrasion resistance of leather, and the shrinkage and color fastness of fabrics, directly affect the quality of footwear and apparel products. If the material characteristics do not meet the requirements or there are quality discrepancies, various quality defects will occur, including cracking at bends due to insufficient leather flexibility, and fading due to poor color fastness of fabrics. The process dimension covers the specific parameters and operating methods of cutting, sewing, ironing and other processes, such as needle length, stitch tension, presser foot pressure, cutting size accuracy, ironing temperature and time, etc. Improper parameter settings or improper operation, such as excessive needle length resulting in loose stitches, inaccurate cutting size resulting in mismatched component sizes, can also cause quality defects. In the reasoning process of shoe and clothing defect location, only the parameters involved in the influence of temperature and humidity on stitch length are considered, which does not mean that these are the only parameters in the reasoning process of shoe and clothing defect location.

[0020] In one embodiment of the present invention, the three-dimensional knowledge graph of S1 also includes relationship edge attributes, and the relationship variable attributes include the influence of temperature and humidity on materials, the needle distance compensation of the process on materials, and the pressure foot force.

[0021] The working principle and effect of the above technical solution are as follows: Temperature and humidity sensors collect temperature and humidity data from the production environment in real time and connect to the material dimension node through a relationship edge. Different leather types react differently to temperature and humidity. Different thicknesses expand or contract to varying degrees under the influence of temperature and humidity, and their surface properties also change due to temperature and humidity. The inference engine analyzes this data and determines the extent to which temperature and humidity affect the material properties. Based on the material properties affected by temperature and humidity, the inference engine searches the knowledge graph for the relationship edge between the material and the theoretical value of the stitch length in the process dimension. The compensation rule library contains stitch length compensation rules under different conditions. The inference engine calculates the stitch length compensation amount and feeds it back to the production equipment or operator. Similarly, the inference engine analyzes the relationship edge between the material properties and the presser foot force node in the process dimension. When temperature and humidity change the material properties, the leather hardness changes, requiring a corresponding adjustment of the presser foot force. The inference engine calculates the appropriate adjusted presser foot force based on the relationship edge rules and displays it in the visual interface. The system adjusts the actual presser foot force based on the real-time presser foot force obtained by the pressure sensor to ensure that the actual and adjusted presser foot forces are consistent. This forms a closed-loop feedback control system, reducing quality defects in footwear and apparel products and improving product quality.

[0022] In one embodiment of the present invention, in addition to acquiring data of each node dimension, S2 also acquires the following data, specifically including: The ambient temperature and humidity are obtained through the temperature and humidity sensor, and the real-time pressure foot force is obtained through the electric pressure sensor.

[0023] The working principle and effect of the above technical solution are as follows: the capacitive humidity sensor uses the capacitance change characteristics of the moisture-sensitive material after absorbing moisture, and the thermistor temperature sensor uses the resistance change characteristics of the temperature to convert the physical quantities of temperature and humidity in the environment into electrical signals and transmit them to the data acquisition system, thereby obtaining ambient temperature and humidity data; the electric pressure sensor is installed on the presser foot part of the shoe and clothing sewing equipment to realize the acquisition of real-time presser foot force data, thereby improving the accuracy of the data.

[0024] In one embodiment of the present invention, the inference engine of S3 further includes a compensation rule library, and according to the rules in the compensation rule library, the needle distance is compensated, specifically including: Determine the compensation method for stitch length according to leather type and humidity; When the leather type is vegetable tanned leather and the humidity is greater than the first compensation threshold, the first compensation method is used for compensation; When the leather type is chrome tanned leather and the humidity is less than the second compensation threshold, the second compensation method is used for compensation.

[0025] The first compensation method includes: when the leather type is vegetable tanned leather and the humidity is greater than the first compensation threshold, the stitch length compensation amount is adjusted in a decreasing direction for every 1% increase in humidity; at the same time, when the temperature exceeds 25°C, the stitch length compensation amount is adjusted in a decreasing direction for every 1°C increase in humidity. The final stitch length compensation amount is the sum of the two parts; The second compensation method includes adjusting the stitch length compensation amount in an increasing direction for every 1% decrease in humidity when the leather type is chrome-tanned leather and the humidity is less than a second compensation threshold.

[0026] The working principle and effect of the above technical solution are as follows: First, the leather type, ambient humidity and temperature data are obtained in real time. If the leather type is detected to be vegetable tanned leather and the humidity is greater than the first compensation threshold (60%), and the temperature exceeds 25°C, the current humidity is set to H% (H>60), and the portion exceeding the humidity threshold is (H-60)%. The current temperature is T°C (T>25), and the portion exceeding the temperature threshold is (T-25)°C. According to the first compensation method, the stitch length compensation amount is reduced by a unit (a = 0.1mm / humidity) for every 1% increase in humidity, and by b units (b = 0.05mm / °C) for every 1°C increase in temperature. The final stitch length compensation amount is mm; among them, Indicates the surface roughness of vegetable tanned leather (0.0032-0.0125mm); Indicates the surface roughness influence coefficient ; If the leather type is chrome tanned leather and the humidity is less than the second compensation threshold (40%), the current humidity is set to H% (H < 40), and the portion below the humidity threshold is (40 - H)%. According to the second compensation method, the stitch length compensation amount is increased by c units (c = 0.15mm) for every 1% decrease in humidity. The stitch length compensation amount is then mm, where The inference engine, which represents the surface roughness of chrome-tanned leather (0.0005-0.002mm), compensates the stitch length in real time based on these calculated stitch length compensation amounts to adapt to the characteristics of different leather types under temperature and humidity changes, thereby optimizing the footwear and clothing sewing process.

[0027] Surface roughness affects needle and thread penetration and friction during sewing. A rougher leather surface increases friction, making sewing more difficult and leading to uneven stitch lengths. By incorporating surface roughness into the stitch length compensation formula, stitch length can be adjusted based on the specific leather surface conditions, minimizing the negative effects of surface roughness differences. Using the compensation rule library to compensate for the needle length improves the accuracy of needle length adaptation and avoids quality defects caused by improper needle length; the needle length is always matched with the leather state to avoid problems such as skewed stitches and cracked seams. At the same time, it improves the production's adaptability to environmental changes and avoids the interference of environmental fluctuations on product quality.

[0028] In one embodiment of the present invention, after the compensation rule library compensates for the stitch length, an adjusted presser foot force is obtained based on the stitch length compensation amount, and the adjusted presser foot force is displayed on a visual interface. The system obtains and displays the real-time presser foot force on the visual interface based on the electric pressure sensor, and adjusts the real-time presser foot force to be consistent with the adjusted presser foot force. Furthermore, the adjusted presser foot force is obtained by the following formula: When the leather type is vegetable tanned leather,

[0029] Among them, F represents the adjusted presser foot force, Indicates the real-time presser foot force, α indicates the needle distance compensation influence coefficient (α=0.09N / mm), Indicates the change in stitch length. Indicates the influence coefficient of leather thickness , t represents the leather thickness (mm), k represents the leather ductility influence coefficient (k=0.002N / mm); Indicates the extension of vegetable tanned leather; Furthermore, the extension of the vegetable tanned leather is obtained by the following formula:

[0030] Where H represents the ambient humidity, Indicates the high humidity elongation coefficient ; Indicates the initial length of vegetable tanned leather in dry state; Indicates the elastic coefficient of leather (when 0<t≤4, (t) = 0.05 + 0.03 (t-2); when t> 4, (t) = 0.1) When the leather type is chrome tanned leather,

[0031] in, Indicates the extension of chrome-tanned leather; Furthermore, the chrome tanned leather extension is obtained by the following formula:

[0032] in, Indicates low humidity shrinkage coefficient .

[0033] The working principle and effect of the above technical solution are as follows: In actual applications, the thicker the material, the greater its elastic modulus and the smaller its ductility. Therefore, the elastic coefficient parameter is introduced into the formula for calculating the ductility of leather. The ductility calculation formula adjusts the elasticity and ductility balance of leather of different thicknesses by quantifying the relationship between thickness and ductility. The accuracy of leather elasticity and ductility prediction avoids the problem of uneven leather quality caused by thickness variations. At the same time, the formula combines leather thickness and ambient humidity. Whether it is a change in thickness or a change in humidity, it will have a corresponding impact on the ductility based on their respective physical effects, and these effects are interrelated. For example, in an environment with high humidity, even if the leather is thicker, its ductility will increase due to the effect of humidity. Compared with the influence of only a single factor on the ductility of leather, the applicability of this formula is improved.

[0034] In the adjusted presser foot force calculation formula, the stitch length reflects the density of the stitches distributed on the leather. Increasing the stitch length means the distance between adjacent stitches widens, requiring a greater presser foot force to ensure the leather remains stable during sewing and prevent displacement during needle penetration and thread pulling, thereby ensuring sewing accuracy and quality. Conversely, decreasing the stitch length increases the density of the stitches, which in turn enhances the leather's holding power, requiring a smaller presser foot force. To ensure smooth needle penetration and even distribution of stitches, a greater presser foot force is required to secure the leather. Thinner leather, however, is relatively soft and less able to withstand external forces. Excessive presser foot force may cause deformation or damage, thus requiring a smaller presser foot force. Therefore, considering the leather thickness parameter allows the presser foot force to be adjusted according to the leather's physical properties, avoiding the impact of improper presser foot force on sewing quality.

[0035] This formula, by quantifying and adding stitch length compensation and leather thickness, accurately calculates presser foot force based on actual stitch length and leather thickness variations during the sewing process. This reduces problems caused by improper presser foot force, including leather deformation, loose seams, and uneven stitch length. Leather's ductility reflects its ability to deform under load, and ambient humidity is a factor that influences leather's ductility. Incorporating this ductility parameter into the calculation formula allows the system to dynamically adjust process parameters, such as presser foot force, based on changes in ambient humidity to adapt to the leather's ductility characteristics at varying humidity levels, thus avoiding sewing defects caused by leather stretch. In one embodiment of the present invention, S3 specifically includes: Acquire real-time data on product quality defects, match the real-time data with the quality defect dimensions in the three-dimensional knowledge graph, confirm the corresponding quality defect nodes, and use the corresponding quality defect nodes as the starting point for reasoning. Based on the preset relationship edges, reverse reasoning is performed along the quality defect dimension, process dimension, material dimension, and environmental dimension to generate a multi-step reasoning path and ultimately determine the cause of the defect.

[0036] The working principle and effect of the above technical solution are as follows: in the S3 stage, the reasoning engine matches the real-time data of quality defects of footwear and clothing products obtained by the sensor with the quality defect dimensions in the three-dimensional knowledge graph, thereby confirming the corresponding quality defect node and using it as the starting point for reasoning. Afterwards, reverse reasoning is performed along the quality defect dimension, process dimension, material dimension and environmental dimension based on the preset relationship edges. During the reasoning process, the relationship between the established dimensions in the knowledge graph is used to gradually trace the causes of quality defects. In the process of reasoning, this method takes into account and avoids the situation where environmental factors cause changes in leather properties, thereby affecting product quality, thereby enhancing the adaptability of the production process to the environment.

[0037] One embodiment of the present invention is a shoe and clothing product defect location system based on multi-step reasoning and knowledge graph, characterized in that the system includes: Multi-dimensional knowledge graph construction system: constructs a three-dimensional knowledge graph with three node dimensions, including material dimension, process dimension and quality defect dimension; Full-dimensional data acquisition system: obtain data of each node dimension; Defect tracing reasoning system: The reasoning engine uses the real-time quality defect data obtained by sensors to reversely reason along the quality defect dimension, process dimension, material dimension, and environmental dimension to generate a multi-step reasoning path and determine the cause of the defect.

[0038] The working principle and effect of the above technical solution are as follows: First, in stage S1, a three-dimensional knowledge graph is constructed, including material, process, and quality defect dimensions. The attributes of the relationship edges are also determined, thereby establishing a correlation framework between the dimensions. Next, in stage S2, data for each node dimension is acquired. Finally, in stage S3, the inference engine matches the real-time sensor-generated quality defect data of footwear and apparel products with the quality defect dimensions in the three-dimensional knowledge graph, identifying the corresponding quality defect node and using it as the starting point for inference. Subsequently, based on the pre-set relationship edges, reverse inference is performed along the quality defect, process, material, and environmental dimensions. During the inference process, the relationships between the dimensions established in the knowledge graph are leveraged to gradually trace the causes of the quality defects. This method considers and avoids the possibility that environmental factors may alter leather properties and thus affect product quality, thereby enhancing the environmental adaptability of the production process.

[0039] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A method for locating defects in footwear and clothing products based on multi-step reasoning and knowledge graph, characterized by: The method comprises: S1: Construct a three-dimensional knowledge graph including three node dimensions, wherein the node dimensions include material dimension, process dimension, and quality defect dimension; S2: Get the data of each node dimension; S3: The reasoning engine uses the real-time quality defect data obtained by sensors to reversely reason along the quality defect dimension, process dimension, material dimension, and environmental dimension to generate a multi-step reasoning path and determine the cause of the defect.

2. The method for locating defects in footwear and clothing products based on multi-step reasoning and knowledge graph according to claim 1, characterized in that: The node dimensions of S1 include the environment dimension, material dimension, process dimension and quality defect dimension, specifically including: Said material dimensions include, leather type, leather thickness and leather surface characteristics; The process dimensions include the theoretical value of needle length and presser foot force; The quality defect dimensions include seam cracking, skewed stitches and uneven stitch length.

3. The method for locating defects in footwear and clothing products based on multi-step reasoning and knowledge graph according to claim 2, characterized in that: The leather types include vegetable tanned leather and chrome tanned leather.

4. The method for locating defects in footwear and clothing products based on multi-step reasoning and knowledge graph according to claim 1, characterized in that: The three-dimensional knowledge graph of S1 also includes relationship edge attributes, and the relationship variable attributes include the impact of temperature and humidity on materials, the process's needle distance compensation for materials, and the pressure foot force.

5. The method for locating defects in footwear and clothing products based on multi-step reasoning and knowledge graph according to claim 1, characterized in that: In addition to obtaining data of each node dimension, S2 also obtains the following data, including: The ambient temperature and humidity are obtained through the temperature and humidity sensor, and the real-time pressure foot force is obtained through the electric pressure sensor.

6. The method for locating defects in footwear and clothing products based on multi-step reasoning and knowledge graph according to claim 1, characterized in that: The reasoning engine of S3 also includes a compensation rule library, and compensates the needle distance according to the rules in the compensation rule library, specifically including: Determine the compensation method for stitch length according to leather type and humidity; When the leather type is vegetable tanned leather and the humidity is greater than the first compensation threshold, the first compensation method is used for compensation; When the leather type is chrome tanned leather and the humidity is less than the second compensation threshold, the second compensation method is used for compensation.

7. The method for locating defects in footwear and clothing products based on multi-step reasoning and knowledge graph according to claim 6, characterized in that: The compensation methods specifically include: The first compensation method includes: when the leather type is vegetable tanned leather and the humidity is greater than the first compensation threshold, the stitch length compensation amount is adjusted in a decreasing direction for every 1% increase in humidity; at the same time, when the temperature exceeds 25°C, the stitch length compensation amount is adjusted in a decreasing direction for every 1°C increase in humidity. The final stitch length compensation amount is the sum of the two parts; The second compensation method includes adjusting the stitch length compensation amount in an increasing direction for every 1% decrease in humidity when the leather type is chrome-tanned leather and the humidity is less than a second compensation threshold.

8. The method for locating defects in footwear and clothing products based on multi-step reasoning and knowledge graph according to claim 6, characterized in that: After the compensation rule library compensates for the stitch length, the adjusted presser foot force is obtained according to the stitch length compensation amount, and the adjusted presser foot force is displayed on the visual interface. The system obtains and displays the real-time presser foot force on the visual interface based on the electric pressure sensor, and adjusts the real-time presser foot force to keep consistent with the adjusted presser foot force.

9. The method for locating defects in footwear and clothing products based on multi-step reasoning and knowledge graph according to claim 1, characterized in that: The S3 specifically includes: Acquire real-time data on product quality defects, match the real-time data with the quality defect dimensions in the three-dimensional knowledge graph, confirm the corresponding quality defect nodes, and use the corresponding quality defect nodes as the starting point for reasoning. Based on the preset relationship edges, reverse reasoning is performed along the quality defect dimension, process dimension, material dimension, and environmental dimension to generate a multi-step reasoning path and ultimately determine the cause of the defect.

10. A shoe and clothing product defect location system based on multi-step reasoning and knowledge graph, characterized by: The system comprises, Multi-dimensional knowledge graph construction system: constructs a three-dimensional knowledge graph with three node dimensions, including material dimension, process dimension and quality defect dimension; Full-dimensional data acquisition system: obtain data of each node dimension; Defect tracing reasoning system: The reasoning engine uses the real-time quality defect data obtained by sensors to reversely reason along the quality defect dimension, process dimension, material dimension, and environmental dimension to generate a multi-step reasoning path and determine the cause of the defect.

Images